GH-39697: [R] Source build should check if offline#39699
GH-39697: [R] Source build should check if offline#39699assignUser merged 5 commits intoapache:mainfrom
Conversation
|
|
|
Macbuilder with just these changes: https://mac.r-project.org/macbuilder/results/1705635740-c74d4358db0b3bdb/ ✅ grabs the binary and uses that
Macbuilder with download.file <- stop() and a release number: https://mac.r-project.org/macbuilder/results/1705638062-f8f78931da423e55/ Macbuilder with download.file <- stop() and a release number and with setting LIBARROW_MINIMAL=TRUE if !download_ok: https://mac.r-project.org/macbuilder/results/1705638501-71bdade97614401d/ ✅ Enables a minimal build: I'm slightly surprised |
|
Hmm the download.file as stop didn't seem to turn off features. |
|
The "and a release number" is needed because Lines 915 to 916 in 55afcf0 But also I think we need to add |
|
And running this locally (with a malignant Build log |
|
Ok, #39699 (comment) is updated |
r/tools/nixlibs.R
Outdated
There was a problem hiding this comment.
If the logic is correct, we shouldn't need this because of https://github.com/apache/arrow/blob/493764929bb2b5635c4d3afff79b2d372495373b/r/tools/nixlibs.R#L585-L599 in build_libarrow()
4937649 to
ab0598e
Compare
|
@github-actions crossbow submit r-binary-packages |
|
Revision: 01eb6ee Submitted crossbow builds: ursacomputing/crossbow @ actions-7e0d11b7c4
|
CRAN. See the commit messages Existing CI should pass and not be affected. We should confirm that source builds get all features enabled. We should test on macbuilder with this package and with one where we've inserted `download.file <- function(...) stop()` at the top of nixlibs.R to simulate offline behavior. I hope there is only one user affected. * Closes: #39697 Authored-by: Neal Richardson <neal.p.richardson@gmail.com> Signed-off-by: Jacob Wujciak-Jens <jacob@wujciak.de>
|
After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 05b8f36. There was 1 benchmark result with an error:
There were no benchmark performance regressions. 🎉 The full Conbench report has more details. It also includes information about 4 possible false positives for unstable benchmarks that are known to sometimes produce them. |
### Rationale for this change CRAN. ### What changes are included in this PR? See the commit messages ### Are these changes tested? Existing CI should pass and not be affected. We should confirm that source builds get all features enabled. We should test on macbuilder with this package and with one where we've inserted `download.file <- function(...) stop()` at the top of nixlibs.R to simulate offline behavior. ### Are there any user-facing changes? I hope there is only one user affected. * Closes: apache#39697 Authored-by: Neal Richardson <neal.p.richardson@gmail.com> Signed-off-by: Jacob Wujciak-Jens <jacob@wujciak.de>
### Rationale for this change CRAN. ### What changes are included in this PR? See the commit messages ### Are these changes tested? Existing CI should pass and not be affected. We should confirm that source builds get all features enabled. We should test on macbuilder with this package and with one where we've inserted `download.file <- function(...) stop()` at the top of nixlibs.R to simulate offline behavior. ### Are there any user-facing changes? I hope there is only one user affected. * Closes: #39697 Authored-by: Neal Richardson <neal.p.richardson@gmail.com> Signed-off-by: Jacob Wujciak-Jens <jacob@wujciak.de>
Rationale for this change
CRAN.
What changes are included in this PR?
See the commit messages
Are these changes tested?
Existing CI should pass and not be affected. We should confirm that source builds get all features enabled. We should test on macbuilder with this package and with one where we've inserted
download.file <- function(...) stop()at the top of nixlibs.R to simulate offline behavior.Are there any user-facing changes?
I hope there is only one user affected.